1. Exposure Curve Pitfalls 2010 Seminar on Reinsurance

2. 2 Antitrust Notice from the CAS The Casualty Actuarial Society is committed to adhering strictly to the letter and spirit of the antitrust laws. Seminars conducted under the auspices of the CAS are designed solely to provide a forum for the expression of various points of view on topics described in the programs or agendas for such meetings. Under no circumstances shall CAS seminars be used as a means for competing companies or firms to reach any understanding – expressed or implied – that restricts competition or in any way impairs the ability of members to exercise independent business judgment regarding matters affecting competition. It is the responsibility of all seminar participants to be aware of antitrust regulations, to prevent any written or verbal discussions that appear to violate these laws, and to adhere in every respect to the CAS antitrust compliance policy.

3. 3 Basic Copyright Notice & Disclaimer for Swiss Re Presentations provided to External Parties ©2009 Swiss Re. All rights reserved. You are not permitted to create any modifications or derivatives of this presentation without the prior written permission of Swiss Re. This presentation is for information purposes only and contains non-binding indications as well as personal judgment. It does not contain any recommendation, advice, solicitation, offer or commitment to effect any transaction or to conclude any legal act. Any opinions or views expressed are of the author and do not necessarily represent those of Swiss Re. Swiss Re makes no warranties or representations as to this presentation’s accuracy, completeness, timeliness or suitability for a particular purpose. Anyone shall at its own risk interpret and employ this presentation without relying on it in isolation. In no event will Swiss Re or one of its affiliates be liable for any loss or damages of any kind, including any direct, indirect or consequential damages, arising out of or in connection with the use of this presentation.

4. 4 Introduction - The goal is to share insights and provide constructive criticism to help you think about improving the curves and their applications. - Although important, we will not focus on Loss Ratio selection. - This is not an endorsement of any specific methodology. Rather, this non-technical presentation is a collection of factual observations and opinions to improve the application of exposure curves in pricing. We need to be mindful of our anti-trust responsibilities. - The remarks included are not a criticism of the organizations that work hard to analyze data and provide the exposure curves. - Any opinions or views expressed are of the author and do not necessarily represent those of Swiss Re.

5. 5 Exposure Curve Pitfalls 5

6. 6 Exposure Curves exclude data from the Excess and Surplus Lines market. - Non admitted data is not reported to ISO. - E&S risks as a percentage of the total market is highly dependent on the cycle. So even if one can accept E&S risks not being in the data, it is not a stable population. - These risks are often unique which sometimes leads to specialized policy forms and/or rates. - It is difficult to know how well default exposure curves describe the true underlying severity distribution of these types of risks. - One needs to understand why a risk is in the E&S market to determine whether it requires an exposure curve adjustment.

7. 7 Exposure Curve Pitfalls Quantity of data excess of a certain threshold drops off. - Most Commercial Auto and GL primary policies have limits of $1M. For personal lines, policy limits are typically smaller. - Several large industrial insurers do not report to ISO. - There may be fewer Liability losses excess of $1M to which curves are fit. - For insight, PSOLD provides occurrence counts excess of certain thresholds that may not be available with other curves. - Pricing common layers of $4m xs $1m and $5m xs $5m are generally based off of extrapolations of curves for Liability. - Are the curves appropriate excess of $10M?

8. 8 Exposure Curve Pitfalls Some Exposure Curves fit to open claims. - Development for WC can be quite long. - When individual claims are developed to ultimate, how good do we feel about the tail factors? Small differences in tail estimation can vary the ultimates quite a bit. This variance difference can have a larger impact on a layer loss cost than a ground up ultimate. - How do we charge for this “parameter uncertainty”? We need to be aware of how open claims are handled in the curve fitting process. - NCCI applies random development which should mitigate this risk for WC curves.

9. 9 Exposure Curve Pitfalls Exposure Curves are often trended to some recent average accident date - If there is a difference in the average accident date in the curves vs the average accident date in the treaty, it could be minor if it is off by only a month or two. - Also, we can trend the curves to a desired average accident date. - However, one needs to consider future loss trend scenarios and make thoughtful trend selections. - For example, if there is a future spike in inflation and losses are sensitive to the date of payment, the exposure curves would require an adjustment to reflect this scenario. The impact could be magnified for an XOL treaty.

10. 10 Exposure Curve Pitfalls Exposure Curves do not vary by policy limit or size of the primary insured - Limited average severities could be a function of some variable like the asset size or revenues of the primary insured. - For Commercial Liability, the curves do not vary by policy limit purchased. However, the Property curves vary by Amount of Insurance. - For insureds that buy higher policy limits, the shape of the exposure curve may be different. - The impact may differ by line of business.

11. 11 Exposure Curve Pitfalls Exposure Curves do not vary for insureds with deductibles/SIRs for Casualty business - This issue is quite similar to whether the exposure curves should vary by size (policy limit) of the insured. - A problem is that insureds who purchase deductibles may behave differently than those who do not. Also, such purchases may vary across the cycle. - Would purchasing a small or large deductible change the loss distribution? For example, an insured may decide to not report a claim as they believe it is within the deductible and so the insurer finds out about the claim late which impacts its severity. - Policy limits and deductibles are probably correlated. Thus, isolating their effects may be difficult.

12. 12 Exposure Curve Pitfalls A frequency shift that varies by size of loss can distort exposure curves - Frequency has generally been negative for many lines of business. - Frequency trend that varies by size of loss in the past or future, could impact average severities. - For example, if frequency trend is negative for small losses but flat or positive for large losses, the exposure curve average severities would be understated. - For WC, the issue might be mitigated since the curves are based upon injury type. However, if there are differences in frequency trend by injury type, the loss weights used to combine the curves will be tilted too heavily towards the smaller injury types.

13. 13 Exposure Rating Pitfalls 13

14. 14 Exposure Rating Pitfalls Exposure rating cannot cover all unforeseen circumstances. - The curves are fit against historical loss data. - Emerging or latent risks may not be in the historical experience which could distort the shape of the curves. Examples include risks like Asbestos, Imported dry wall, etc. - A Casualty CAT load could be considered. This could be a frequency or severity issue. Such a load is generally positive but it could be negative. It needs to be thought of in conjunction with the loss ratio selection. - Will the inclusion of extreme events make the exposure curves more or less severe? Could a change in the curve shape affect lower layers as well?

15. 15 Exposure Rating Pitfalls Exposure Curves are not designed for Clash or ECO/XPL losses. - The curves are not fit against occurrences which include the sum of losses from multiple lines of business or coverage. - Reinsurance treaties usually work on an occurrence basis. Per event covers can also present a challenge. - For a treaty that covers WC and Commercial Auto, the curves will not properly charge for an occurrence that includes both WC and Commercial Auto losses. - Consider modeling for Clash or ECO/XPL exposure.

16. 16 Exposure Rating Pitfalls Exposure Curves will not include a charge for Multi Peril Losses - A treaty covering Liability and Property risks may have a basket retention. - The exposure curves are not designed for occurrences with both Property and Liability losses. - Assume the treaty covers Commercial Multi Peril policies. If there is a fire on the premises and people are injured, the curves will not properly charge for such an occurrence. - A charge may need to be considered for Multi Peril basket exposure.

17. 17 Exposure Rating Pitfalls Challenges with modeling ALAE - ALAE from the Increased Limits Factor circulars may not be appropriate for XOL pricing. - How can the curves be modified to handle ALAE included? - Should ALAE vary by size of loss? - If ALAE is unlimited in the primary policy but included as part of the loss in the XOL treaty, how does one price for higher layers that are unexposed by policy limits but exposed by unlimited ALAE? - Should a random or probabilistic ALAE approach be used? - ISO does have a joint loss and ALAE distribution.

18. 18 Exposure Rating Pitfalls Exposure curves are not perfectly designed for individual risks - When experience rating lacks credible loss activity, one tends to put greater weight on exposure rating which may or may not be valid. - Just because experience is not credible, that does not mean that an exposure rating is credible (or vice versa). - However, each risk really has its own unique severity distribution based upon such things like the insured’s operations, location, legal environment, economic environment, risk management, etc. - How can one consider specific risk drivers and loss scenarios to modify exposure rating results?

19. 19 Exposure Rating Pitfalls Exposure rating for Personal Auto is a challenge - Reinsurance contracts generally work on an occurrence basis. Thus, all Liability losses from an auto accident must be summed and applied to the reinsurance limit and retention. - One cannot use curves that are based solely on Bodily Injury as the Property Damage and Medical Components need to be considered. - UM/UIM stacking and PIP laws vary greatly by state. - Some treaties cover Auto Physical Damage losses as a part of the total loss amount.

20. 20 Exposure Rating Pitfalls Property exposure rating challenges - How accurate are reported Amounts of Insurance? If there is serious underreporting or shifts in underreporting, how will that influence the exposure curves and their accuracy? - Business Interruption is typically covered, but it can be unlimited. - “State” may be a good variable. But, how about other variables such as urban vs rural, economic environment, height of buildings, distance to ocean, etc for non CAT perils? - Separating CAT and non CAT losses can be a challenge. Experience rating should generally exclude CAT losses. Cat exposure rating is often based on vendor CAT models, exposure curves, or judgment. The perils underlying the exposure rating curves need to be consistent with how they are being applied.

21. Thank you – Are there any questions?